Help for package tvrmst

Type:

Package

Title:

Time-Varying Restricted Mean Survival Time from Survival Matrices

Version:

0.0.6

Author:

Imad EL BADISY [aut, cre]

Maintainer:

Imad EL BADISY <elbadisyimad@gmail.com>

Description:

Utilities for restricted mean survival time (RMST) and time-varying restricted mean survival time quantities computed from survival curves provided on a time grid. The package is model-agnostic and accepts only a time vector and survival matrices, returning RMST-based quantities and bootstrap summaries. For restricted mean survival time methodology, see Royston and Parmar (2013) <doi:10.1186/1471-2288-13-152>.

License:

MIT + file LICENSE

Encoding:

UTF-8

RoxygenNote:

7.3.3

Imports:

stats

Suggests:

ggplot2, survival, testthat (≥ 3.0.0), roxygen2 (≥ 7.0.0)

Config/testthat/edition:

NeedsCompilation:

Packaged:

2026-03-11 15:19:35 UTC; imad-el-badisy

Repository:

CRAN

Date/Publication:

2026-03-16 19:40:02 UTC

tvrmst: Time-Varying Restricted Mean Survival Time from Survival Matrices

Description

Author(s)

Maintainer: Imad EL BADISY elbadisyimad@gmail.com

Construct a survival matrix object

Description

Construct a survival matrix object

Usage

as_survmat(S, time, id = NULL, group = NULL)

Arguments

S

Numeric matrix of survival probabilities (rows = subjects, columns = time).

time

Numeric strictly increasing time grid with length(time) == ncol(S).

id

Optional subject identifiers. Defaults to seq_len(nrow(S)).

group

Optional group labels with length nrow(S).

Value

An object of class "survmat".

Examples

time <- c(0, 1, 2)
S <- rbind(
  c(1.0, 0.8, 0.6),
  c(1.0, 0.7, 0.5)
)
x <- as_survmat(S, time, group = c("A", "B"))
x

Coerce survival predictions to a numeric matrix

Description

Coerce survival predictions to a numeric matrix

Usage

as_survprob_matrix(pred, times = NULL)

Arguments

pred

Prediction output from a model.

times

Unused placeholder for API compatibility.

Value

Numeric matrix with rows as subjects and columns as times.

Examples

pred <- data.frame(
  t0 = c(1.0, 1.0),
  t1 = c(0.8, 0.7),
  t2 = c(0.6, 0.5)
)
as_survprob_matrix(pred)

Row-bind survmat objects on a common time grid

Description

Row-bind survmat objects on a common time grid

Usage

bind_survmat(...)

Arguments

...

One or more survmat objects.

Value

A combined survmat.

Examples

time <- c(0, 1, 2)
xA <- as_survmat(matrix(c(1.0, 0.9, 0.7), nrow = 1), time, group = "A")
xB <- as_survmat(matrix(c(1.0, 0.8, 0.6), nrow = 1), time, group = "B")
x <- bind_survmat(xA, xB)
nobs_survmat(x)

Bootstrap CI for dynamic RMST delta curve

Description

Bootstrap CI for dynamic RMST delta curve

Usage

boot_rmst_delta(xA, xB, R = 300, conf = 0.95, seed = NULL)

Arguments

xA

A survmat object for arm A.

xB

A survmat object for arm B.

R

Number of bootstrap replicates.

conf

Confidence level in ⁠(0, 1)⁠.

seed

Optional RNG seed.

Value

A list with point estimate and percentile confidence bands.

Examples

time <- c(0, 1, 2)
xA <- as_survmat(rbind(c(1.0, 0.9, 0.7), c(1.0, 0.8, 0.6)), time, group = c("A", "A"))
xB <- as_survmat(rbind(c(1.0, 0.95, 0.8), c(1.0, 0.85, 0.7)), time, group = c("B", "B"))
boot <- boot_rmst_delta(xA, xB, R = 10, seed = 1)
boot$estimate

Generic bootstrap for curve-valued estimators

Description

Generic bootstrap for curve-valued estimators

Usage

bootstrap_curve(est_fun, R = 300, conf = 0.95, seed = NULL, keep_reps = FALSE)

Arguments

est_fun

Function of one argument r; called with NULL for the point estimate and with integers 1:R for bootstrap replicates.

R

Number of replicates.

conf

Confidence level in ⁠(0, 1)⁠.

seed

Optional RNG seed.

keep_reps

If TRUE, include replicate matrix in output.

Value

A list with estimate, lo, hi, and conf.

Examples

vals <- c(1, 2, 4, 8)
boot <- bootstrap_curve(
  function(r) {
    if (is.null(r)) {
      mean(vals)
    } else {
      mean(sample(vals, replace = TRUE))
    }
  },
  R = 20,
  seed = 1
)
boot$estimate

Number of rows in a survmat

Description

Number of rows in a survmat

Usage

nobs_survmat(x)

Arguments

x

A survmat object.

Value

Number of subjects.

Examples

time <- c(0, 1, 2)
S <- rbind(
  c(1.0, 0.8, 0.6),
  c(1.0, 0.7, 0.5)
)
x <- as_survmat(S, time)
nobs_survmat(x)

Plot bootstrap estimate with confidence ribbon

Description

Plot bootstrap estimate with confidence ribbon

Usage

plot_boot_curve(
  boot,
  grid = NULL,
  title = "Bootstrap curve",
  xlab = "t",
  ylab = "estimate",
  x_scale = 1,
  y_scale = 1,
  x_unit = NULL,
  y_unit = NULL
)

Arguments

boot

List returned by bootstrap_curve() or boot_rmst_delta().

grid

Optional x-axis grid. If omitted, uses boot$time.

title

Plot title.

xlab

X-axis label.

ylab

Y-axis label.

x_scale

Positive divisor applied to the x-axis values before plotting.

y_scale

Positive divisor applied to the y-axis values before plotting.

x_unit

Optional unit label appended to the x-axis title.

y_unit

Optional unit label appended to the y-axis title.

Value

A ggplot object.

Examples


time <- c(0, 1, 2)
xA <- as_survmat(rbind(c(1.0, 0.9, 0.7), c(1.0, 0.8, 0.6)), time, group = c("A", "A"))
xB <- as_survmat(rbind(c(1.0, 0.95, 0.8), c(1.0, 0.85, 0.7)), time, group = c("B", "B"))
boot <- boot_rmst_delta(xA, xB, R = 10, seed = 1)
plot_boot_curve(boot)

Plot a delta curve

Description

Plot a delta curve

Usage

plot_delta_curve(
  grid,
  delta,
  title = "Delta curve",
  xlab = "t",
  ylab = "Delta",
  x_scale = 1,
  y_scale = 1,
  x_unit = NULL,
  y_unit = NULL
)

Arguments

grid

X-axis grid.

delta

Y values.

title

Plot title.

xlab

X-axis label.

ylab

Y-axis label.

x_scale

Positive divisor applied to the x-axis values before plotting.

y_scale

Positive divisor applied to the y-axis values before plotting.

x_unit

Optional unit label appended to the x-axis title.

y_unit

Optional unit label appended to the y-axis title.

Value

A ggplot object.

Examples


time <- c(0, 1, 2)
xA <- as_survmat(rbind(c(1.0, 0.9, 0.7), c(1.0, 0.8, 0.6)), time, group = c("A", "A"))
xB <- as_survmat(rbind(c(1.0, 0.95, 0.8), c(1.0, 0.85, 0.7)), time, group = c("B", "B"))
d <- rmst_delta(xA, xB)
plot_delta_curve(d$time, d$delta)

Plot individual dynamic RMST curves by group

Description

Plot individual dynamic RMST curves by group

Usage

plot_rmst_individual_by_group(
  res,
  group,
  n_show_per_group = 30,
  title = "Individual dynamic RMST by group",
  x_scale = 1,
  y_scale = 1,
  x_unit = NULL,
  y_unit = NULL
)

Arguments

res

Result from rmst_dynamic() containing individual and time.

group

Group vector aligned with rows of res$individual.

n_show_per_group

Maximum number of individual curves shown per group.

title

Plot title.

x_scale

Positive divisor applied to the x-axis values before plotting.

y_scale

Positive divisor applied to the y-axis values before plotting.

x_unit

Optional unit label appended to the x-axis title.

y_unit

Optional unit label appended to the y-axis title.

Value

A ggplot object.

Examples


time <- c(0, 1, 2)
S <- rbind(
  c(1.0, 0.9, 0.7),
  c(1.0, 0.8, 0.6),
  c(1.0, 0.95, 0.8),
  c(1.0, 0.85, 0.7)
)
grp <- c("A", "A", "B", "B")
x <- as_survmat(S, time, group = grp)
res <- rmst_dynamic(x)
plot_rmst_individual_by_group(res, grp, n_show_per_group = 2)

Plot mean RMST curves for two arms

Description

Plot mean RMST curves for two arms

Usage

plot_rmst_two_arms(
  xA,
  xB,
  labels = c("Arm A", "Arm B"),
  title = NULL,
  xlab = "Time (tau)",
  ylab = "RMST(tau)",
  curve_colors = c("#1B6CA8", "#D95F02"),
  x_scale = 1,
  y_scale = 1,
  x_unit = NULL,
  y_unit = NULL
)

Arguments

xA

A survmat object for arm A.

xB

A survmat object for arm B.

labels

Two legend labels.

title

Plot title. Use NULL to omit it.

xlab

X-axis label.

ylab

Y-axis label.

curve_colors

Colors for the two curves.

x_scale

Positive divisor applied to the x-axis values before plotting.

y_scale

Positive divisor applied to the y-axis values before plotting.

x_unit

Optional unit label appended to the x-axis title.

y_unit

Optional unit label appended to the y-axis title.

Value

A ggplot object.

Examples


time <- c(0, 1, 2)
xA <- as_survmat(rbind(c(1.0, 0.9, 0.7), c(1.0, 0.8, 0.6)), time, group = c("A", "A"))
xB <- as_survmat(rbind(c(1.0, 0.95, 0.8), c(1.0, 0.85, 0.7)), time, group = c("B", "B"))
plot_rmst_two_arms(xA, xB, labels = c("Control", "Treatment"))

Dynamic RMST difference between two arms

Description

Computes Delta(tau) = RMST_B(tau) - RMST_A(tau).

Usage

rmst_delta(xA, xB, tau = NULL)

Arguments

xA

A survmat object for arm A.

xB

A survmat object for arm B.

tau

Optional numeric horizons for interpolation.

Value

A list with RMST curves for both arms and the delta curve.

Examples

time <- c(0, 1, 2)
xA <- as_survmat(rbind(c(1.0, 0.9, 0.7), c(1.0, 0.8, 0.6)), time, group = c("A", "A"))
xB <- as_survmat(rbind(c(1.0, 0.95, 0.8), c(1.0, 0.85, 0.7)), time, group = c("B", "B"))
rmst_delta(xA, xB, tau = 1.5)$at_tau

Dynamic RMST trajectories

Description

Computes ⁠RMST_i(tau) = integral_0^tau S_i(u) du⁠ over the provided grid.

Usage

rmst_dynamic(x, tau = NULL, by = x$group)

Arguments

x

A survmat object.

tau

Optional numeric horizons for interpolation.

by

Optional grouping variable. Defaults to x$group.

Value

A list with individual and mean RMST curves.

Examples

time <- c(0, 1, 2)
S <- rbind(
  c(1.0, 0.8, 0.6),
  c(1.0, 0.7, 0.5)
)
x <- as_survmat(S, time, group = c("A", "A"))
res <- rmst_dynamic(x, tau = c(0.5, 1.5))
res$mean
res$at_tau

tvrmst: Matrix-first dynamic restricted mean survival time utilities

Description

Tools for dynamic restricted mean survival time estimation, restricted mean survival time contrasts, bootstrap uncertainty, and plotting from survival probability matrices evaluated on a common time grid.